CN108490000A - A kind of Bar Wire Product surface defect on-line measuring device and method - Google Patents
A kind of Bar Wire Product surface defect on-line measuring device and method Download PDFInfo
- Publication number
- CN108490000A CN108490000A CN201810205229.7A CN201810205229A CN108490000A CN 108490000 A CN108490000 A CN 108490000A CN 201810205229 A CN201810205229 A CN 201810205229A CN 108490000 A CN108490000 A CN 108490000A
- Authority
- CN
- China
- Prior art keywords
- image
- wire product
- bar wire
- red
- green
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8806—Specially adapted optical and illumination features
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
- G01N21/952—Inspecting the exterior surface of cylindrical bodies or wires
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Chemical & Material Sciences (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Analytical Chemistry (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The present invention relates to a kind of device and method of Bar Wire Product surface defect on-line checking, using the image collecting device of multispectral camera and red, green, blue, near-infrared light source combination, it is synchronous to obtain Bar Wire Product surface to be detected gray level image and depth image, and pass through the surface defect of gray level image and depth image fusion method detection Bar Wire Product.Multispectral camera acquires red, green, blue and near-infrared light source in the reflected light on Bar Wire Product surface, is isolated from the image collected red（R）, it is green（G）, it is blue（B）And near-infrared（NIR）R, G, channel B image three-dimensional reconstruction algorithm are rebuild three-dimensional surface, and obtain depth image by channel image using NIR channel images as gray level image.By the Pixel-level fusion detection Bar Wire Product surface defect areas of gray level image and depth image, the gray level image of defect area and depth image are input to convolutional neural networks and carry out defect classification, obtains defect recognition result.
Description
Technical field
The present invention relates to a kind of methods and reality carrying out Bar Wire Product surface defect on-line checking using multi-optical spectrum imaging technology
Existing device, belongs to field of optical measurements.Stick line is obtained simultaneously by the combination of multispectral camera and red, green, blue, near-infrared light source
The gray level image and depth image on material surface pass through gray level image and depth image fusion method on-line checking and identification Bar Wire Product
Surface defect improves the accuracy rate of defects detection and identification.
Background technology
Bar Wire Product surface defect is an important factor for influencing Bar Wire Product quality, at present for the detection of Bar Wire Product surface defect
Big to mostly use the conventional lossless detection method such as ultrasonic, infrared, magnetic leakage, vortex, these methods are only used for offline or low speed feelings
Condition, and the defect type detected is very limited.With the development of CCD camera shooting and computer image processing technology, it is based on machine
The surface detection technique of device vision is of increased attention, and still, current surface detecting system mostly uses greatly with band
The flat steel products such as steel, cut deal are the bidimensional detection method for detecting object, this bidimensional detection mode difficulty in the application of Bar Wire Product
With the effect obtained, the reason is as follows that：
(1) section of Bar Wire Product is cambered surface, if using two dimensional image detection mode, can be existed in the image collected
Shade and uneven illumination phenomenon increase difficulty for subsequent image procossing.
(2) Bar Wire Product surface is covered with a large amount of iron scales, iron scale in appearance without evident regularity, and with stick line
Material surface backfin, impression the real defects such as are mingled with without significant difference, it is difficult to distinguish.
Bar Wire Product surface defect can from the appearance be divided into two classes：One kind is that backfin, impression, pit etc. in appearance can be with
Seeing has the defect of change in depth；It is another kind of be folding, crackle, the defect for being difficult to see that change in depth in appearance such as be mingled with.
The defect of change in depth can be detected using 3 D detection method, and since iron scale is almost without change in depth, because
This can remove iron scale interference by 3 D detection method, but can not detect the defect of not change in depth.If will
Bidimensional detection is combined with three dimensional detection, then can the defect of change in depth be detected the presence of simultaneously, and removes iron scale
Interference, improve Detection accuracy.But bidimensional detection is combined with three dimensional detection and is examined applied to Bar Wire Product surface defect
It surveys, needs to solve the problems, such as 3 following：
(1) three dimensional detection detects the stationary problem in detection resolution and detection speed with bidimensional.Due to three dimensional detection
Data volume it is big, therefore there are the contradictory problems of detection resolution and detection speed, it is difficult to reach and detect comparable point with bidimensional
Resolution and speed just cannot achieve the mutual supplement with each other's advantages of the two if two methods cannot synchronize.
(2) shade in two dimensional image and uneven illumination phenomenon.It needs to reduce shade and uneven illumination phenomenon, to improve
Picture quality when bidimensional detects.
(3) half-tone information and depth information merge problem.Traditional bidimensional is detected from three dimensional detection by different cameras
It realizes, the corresponding image that different cameral obtains is not the same area, needs to be registrated image, increases information fusion
Difficulty.
The present invention will be solved the above problems using multi-optical spectrum imaging technology, by multispectral camera acquisition red, green, blue, closely
Infrared light supply isolates red, green, blue, close in the reflected light on Bar Wire Product surface from the multispectral image that multispectral camera acquires
Infrared channel image obtains surface depth image, near infrared channels figure by the restructural three-dimensional surface of red, green, blue channel image
As can be used as surface gray level image, therefore, a camera can obtain surface depth image and gray level image simultaneously, and pass through depth
Image and surface defect gray level image fusion method detection and identify Bar Wire Product.
Invention content
A kind of Bar Wire Product surface defect on-line measuring device, using multispectral camera and red, green, blue, near-infrared light source group
The image collecting device of conjunction, the multispectral camera have four kinds of sensors of red, green, blue and near-infrared, respectively to red, green, blue,
Near infrared light is sensitive, and the red, green, blue and near-infrared light source all use LED monochromatic light, are evenly distributed on the four of multispectral camera
Week, and it is irradiated to Bar Wire Product surface the same area.The image collecting device that the red, green, blue, near-infrared light source combine needs 4 altogether
Platform, is separately mounted to upper left, upper right, bottom right, the lower left of Bar Wire Product to be detected, and the light path of the multispectral camera passes through stick line
The axle center of material, and with the vertical line in axle center angle at 45 °.Multispectral camera pickup area is identical as light source irradiation area, the region
The pickup areas of perimeter 1/4,4 image collecting devices being more than Bar Wire Product circumference circle can cover the entire circle of Bar Wire Product
Week.
Using a kind of Bar Wire Product surface defect online test method described in claim 1, the multispectral camera is adopted
Collect red, green, blue and near-infrared light source in the reflected light on Bar Wire Product surface, obtains the mostly light for including red, green, blue and near infrared channels
Spectrogram picture isolates red, green, blue and near infrared channels image from multispectral image, using near infrared channels image as stick line
The gray level image on material surface irradiates red, green, blue channel image as Bar Wire Product surface the same area in different lighting angles
Image obtains surface depth image by three-dimensional reconstruction algorithm, is merged with the Pixel-level of depth image by surface gray level image
The gray level image of defect area and depth image are input to convolutional neural networks and lacked by test bar wire surface defect area
Classification and Identification is fallen into, defect recognition result is obtained.
The specific method is as follows：
According to photometric stereo (Photometric Stereo) principle, for the light source that luminous intensity is E, reflected light
Light intensity be：
I=ρ Eln (I)
In formula (1), n is the unit normal vector at certain point on surface；ρ is diffusing reflection coefficient；L=(lx ly lz), it is light source
Unit direction vector；E is incident intensity.
ρR, ρG, ρBFor the diffusing reflection coefficient of red, green, blue light source, ER, EG, EBFor the incident intensity of red, green, blue light source, lead to
Debugging light source is crossed, ρ is madeRER=ρGEG=ρBEB=ρ E.Then had according to formula (1)：
In formula (2), IR,IG,IBThe brightness of respectively R, G, channel B image, lR,lG,lBRespectively red, green, blue light source enters
Penetrate the direction vector of light.
If light source direction matrix isReflective light intensity is：IRGB=(IR IG IB)T, then surface
Normal vector is：
N=L-1·IRGB/|L-1·IRGB| (3)
If the depth that certain in image coordinate system (x, y) is put is Z (x, y), defines it and distinguish in the x-direction with the gradient in the directions y
For P (x, y), Q (x, y) then has：
Surface unit normal vector n is defined in x, y, the component on the directions z is：Nx、Ny、Nz, can be obtained according to formula (4)：
(P, Q)=(Nx./Nz,Ny./Nz) (5)
If the initial value Z of Z0(x, y)=0 can find out the value of Z (x, y) according to iterative formula below：
ZmOptimal solution can be obtained by successive ignition in (x, y).
On the case depth data projection to the plane of delineation that formula (6) is obtained, surface depth image is obtained.Since NIR is logical
It is the same area on Bar Wire Product surface that the depth image that the gray level image and preceding step that road obtains obtain is corresponding, therefore
The Pixel-level fusion that image can be directly carried out without image registration, so as to more accurately detect the area where defect
Domain.The gray level image of defect area and depth image are input to convolutional neural networks and carry out defect Classification and Identification, convolutional Neural
Network has carried out off-line learning with defect sample, finally obtains the recognition result of defect.
The present invention is implemented in combination with Bar Wire Product surface defects detection by bidimensional detection and three dimensional detection, and advantageous effect embodies
：
(1) present invention can obtain Bar Wire Product surface gray level image and depth image simultaneously by same multispectral camera,
Therefore, gray level image and depth image are synchronous in resolution ratio and picking rate, which solves bidimensional detection with it is three-dimensional
Detect the stationary problem in resolution ratio and speed.
(2) the NIR channel images detached from multispectral camera are the infrared imagings on Bar Wire Product surface, with visible images
It compares, the even phenomenon of shade and uneven illumination in NIR images is greatly reduced, to improve the quality of gray level image.
(3) surface gray level image is obtained with depth image by same multispectral camera, and corresponding is the same of Bar Wire Product
Region can realize image co-registration without image registration.
Description of the drawings
Fig. 1 is the schematic diagram of described image harvester, in Fig. 1：1 is multispectral camera, and 2 be red light source, and 3 be green
Light source, 4 be blue-light source, and 5 be near-infrared light source, and 6 be Bar Wire Product to be detected, and 7 adopt for the image on 6 surface of Bar Wire Product to be detected
Collect region.
Fig. 2 is the schematic diagram that described image harvester is circumferentially installed in Bar Wire Product.In Fig. 2：1,2,3,4 be Image Acquisition
Device separately includes multispectral camera 1a, 2a, 3a, 4a, and red light source 1b, 2b, 3b, 4b, green light source 1c, 2c, 3c, 4c are blue
Color light source 1d, 2d, 3d, 4d and near-infrared light source 1e, 2e, 3e, 4e, 5 be Bar Wire Product to be detected.
Fig. 3 is the multi-spectral image processing flow chart of multispectral camera acquisition.
Specific implementation mode
In Fig. 1, the multispectral camera 1 has four kinds of sensors of red, green, blue and near-infrared, respectively to red, green, blue, close
Infrared photaesthesia.The red light source 2, green light source 3, blue-light source 4 and near-infrared light source 5 are LED monochromatic sources, are shone
Wavelength is equal with the corresponding spectral response curve peak value of four kinds of sensors of multispectral camera.The light path of multispectral camera 1 passes through
The axle center of Bar Wire Product 6 is crossed, red light source 2, green light source 3, blue-light source 4 and near-infrared light source 5 are evenly distributed on multispectral phase
The different wavelengths of light of the surrounding of machine 1, light source transmitting is irradiated to the same area 7 on 6 surface of Bar Wire Product, and the perimeter in region 7 is more than stick
The 1/4 of 6 circumference circle of wire rod.Multispectral camera 1 acquires red light source 2, green light source 3, blue-light source 4 and near-infrared light source 5
Reflected light in region 7, obtains multispectral image, with for further processing.
In Fig. 2, image collecting device 1, image collecting device 2, image collecting device 3 and image collecting device 4 are pacified respectively
Mounted in the upper left of Bar Wire Product, upper right, bottom right, lower left, multispectral camera 1a, multispectral camera 2a, multispectral camera 3a and more
The light path of spectrum camera 4a pass through Bar Wire Product axle center, and with the vertical line in axle center angle at 45 °.Due to the figure of image collecting device
As pickup area covering Bar Wire Product circumference is more than 1/4 perimeter, therefore image collecting device 1, image collecting device 2, image are adopted
The image acquisition region of acquisition means 3 and image collecting device 4 can cover the whole circumference of Bar Wire Product 5.
Fig. 3 is the collected multi-spectral image processing flow of multispectral camera, the specific steps are：
(1) red (R), green (G), blue (B) and near-infrared (NIR) channel image are isolated from multispectral image；
(2) using NIR channel images as the gray level image of Rail Surface；
(3) image for irradiating R, G, channel B image in different lighting angles as Bar Wire Product surface the same area, passes through
Formula (6) obtains case depth data；
(4) by the case depth data projection to the plane of delineation obtained in step 3, surface depth image is obtained；
(5) the corresponding depth image that the gray level image and step 4 obtained due to step 2 is obtained is the same of Bar Wire Product surface
One region can directly carry out the Pixel-level fusion of image without image registration；
(6) segmentation for carrying out image and Defect Edge extraction are merged with the Pixel-level of gray level image by depth image, from
And it can more accurately detect the region where defect；
(7) gray level image of defect area and depth image are input to convolutional neural networks and carry out defect Classification and Identification,
Convolutional neural networks have carried out off-line learning with defect sample；
(8) recognition result of defect is finally obtained.
Claims (2)
1. a kind of Bar Wire Product surface defect on-line measuring device, it is characterised in that：Using multispectral camera and red, green, blue, close red
The image collecting device of outer light source combination, the multispectral camera has four kinds of sensors of red, green, blue and near-infrared, right respectively
Red, green, blue, near infrared light are sensitive, and the red, green, blue and near-infrared light source all use LED monochromatic light, are evenly distributed on mostly light
The surrounding of camera is composed, and is irradiated to Bar Wire Product surface the same area, the Image Acquisition that the red, green, blue, near-infrared light source combine
Device needs 4 altogether, is separately mounted to upper left, upper right, bottom right, the lower left of Bar Wire Product to be detected, the light of the multispectral camera
Road pass through Bar Wire Product axle center, and with the vertical line in axle center angle at 45 °, multispectral camera pickup area and light source irradiation area phase
Together, the perimeter in the region can cover Bar Wire Product more than the pickup area of 1/4,4 image collecting devices of Bar Wire Product circumference circle
Whole circumference.
2. utilizing a kind of Bar Wire Product surface defect online test method described in claim 1, it is characterised in that：The mostly light
Camera acquisition red, green, blue and near-infrared light source are composed in the reflected light on Bar Wire Product surface, is obtained logical comprising red, green, blue and near-infrared
The multispectral image in road isolates red, green, blue and near infrared channels image from multispectral image, by near infrared channels image
As the gray level image on Bar Wire Product surface, using red, green, blue channel image as Bar Wire Product surface the same area at different illumination angles
The image for spending irradiation, obtains surface depth image by three-dimensional reconstruction algorithm, passes through the picture of surface gray level image and depth image
Plain grade fusion detection Bar Wire Product surface defect areas, convolutional Neural net is input to by the gray level image of defect area and depth image
Network carries out defect Classification and Identification, obtains defect recognition result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810205229.7A CN108490000A (en) | 2018-03-13 | 2018-03-13 | A kind of Bar Wire Product surface defect on-line measuring device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810205229.7A CN108490000A (en) | 2018-03-13 | 2018-03-13 | A kind of Bar Wire Product surface defect on-line measuring device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108490000A true CN108490000A (en) | 2018-09-04 |
Family
ID=63338993
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810205229.7A Pending CN108490000A (en) | 2018-03-13 | 2018-03-13 | A kind of Bar Wire Product surface defect on-line measuring device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108490000A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109521022A (en) * | 2019-01-23 | 2019-03-26 | 苏州鼎纳自动化技术有限公司 | Touch screen defect detecting device based on the confocal camera of line |
| CN110346294A (en) * | 2019-06-17 | 2019-10-18 | 北京科技大学 | A kind of subtle scanning-detecting system and method for scratching defect of panel |
| CN110813798A (en) * | 2019-12-07 | 2020-02-21 | 浙江科技学院 | Bar defect detection device and method based on vision |
| CN111272763A (en) * | 2018-12-04 | 2020-06-12 | 通用电气公司 | System and method for workpiece inspection |
| CN111650212A (en) * | 2020-07-03 | 2020-09-11 | 东北大学 | Metal surface normal direction three-dimensional information acquisition method based on linear array camera three-dimensional vision |
| CN111811432A (en) * | 2020-06-16 | 2020-10-23 | 中国民用航空飞行学院 | Three-dimensional imaging system and method |
| CN111968094A (en) * | 2020-08-18 | 2020-11-20 | 创新奇智(西安)科技有限公司 | Rod defect detection method and device, electronic equipment and readable storage medium |
| CN112614088A (en) * | 2020-12-01 | 2021-04-06 | 安徽维德工业自动化有限公司 | Identification and detection method based on 3D visual detection technology |
| CN113096110A (en) * | 2021-01-15 | 2021-07-09 | 深圳锦绣创视科技有限公司 | Defect self-detection method based on deep learning and related device |
| CN113146368A (en) * | 2020-05-29 | 2021-07-23 | 浙江大学 | Steel rail surface quality detection system used on long trajectory |
| CN113252567A (en) * | 2021-06-08 | 2021-08-13 | 菲特（天津）检测技术有限公司 | Method, system, medium and terminal for rapidly detecting multiple defects on surface of aluminum valve plate |
| CN114486732A (en) * | 2021-12-30 | 2022-05-13 | 武汉光谷卓越科技股份有限公司 | Ceramic tile defect online detection method based on line scanning three-dimension |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080063426A1 (en) * | 2002-12-03 | 2008-03-13 | Tzyy-Shuh Chang | Apparatus and method for detecting surface defects on a workpiece such as a rolled/drawn metal bar |
| US20110279656A1 (en) * | 2009-05-21 | 2011-11-17 | Yeditepe Universitesi | Surface Scanning System |
| CN102288613A (en) * | 2011-05-11 | 2011-12-21 | 北京科技大学 | Surface defect detecting method for fusing grey and depth information |
| CN103913465A (en) * | 2014-04-04 | 2014-07-09 | 北京科技大学 | Multi-information fused online high-temperature casting blank surface defect detection method |
-
2018
- 2018-03-13 CN CN201810205229.7A patent/CN108490000A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080063426A1 (en) * | 2002-12-03 | 2008-03-13 | Tzyy-Shuh Chang | Apparatus and method for detecting surface defects on a workpiece such as a rolled/drawn metal bar |
| US20110279656A1 (en) * | 2009-05-21 | 2011-11-17 | Yeditepe Universitesi | Surface Scanning System |
| CN102288613A (en) * | 2011-05-11 | 2011-12-21 | 北京科技大学 | Surface defect detecting method for fusing grey and depth information |
| CN103913465A (en) * | 2014-04-04 | 2014-07-09 | 北京科技大学 | Multi-information fused online high-temperature casting blank surface defect detection method |
Non-Patent Citations (2)
| Title |
|---|
| 张子恕 等: "《影像式线棒材在线表面检测系统》", 《轧钢》 * |
| 褚翔 等: ""基于可见光及近红外的仿生视觉成像技术"", 《光电技术应用》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111272763A (en) * | 2018-12-04 | 2020-06-12 | 通用电气公司 | System and method for workpiece inspection |
| CN111272763B (en) * | 2018-12-04 | 2023-04-07 | 通用电气公司 | System and method for workpiece inspection |
| CN109521022A (en) * | 2019-01-23 | 2019-03-26 | 苏州鼎纳自动化技术有限公司 | Touch screen defect detecting device based on the confocal camera of line |
| CN110346294A (en) * | 2019-06-17 | 2019-10-18 | 北京科技大学 | A kind of subtle scanning-detecting system and method for scratching defect of panel |
| CN110813798A (en) * | 2019-12-07 | 2020-02-21 | 浙江科技学院 | Bar defect detection device and method based on vision |
| CN113146368A (en) * | 2020-05-29 | 2021-07-23 | 浙江大学 | Steel rail surface quality detection system used on long trajectory |
| CN111811432A (en) * | 2020-06-16 | 2020-10-23 | 中国民用航空飞行学院 | Three-dimensional imaging system and method |
| CN111650212A (en) * | 2020-07-03 | 2020-09-11 | 东北大学 | Metal surface normal direction three-dimensional information acquisition method based on linear array camera three-dimensional vision |
| CN111968094A (en) * | 2020-08-18 | 2020-11-20 | 创新奇智(西安)科技有限公司 | Rod defect detection method and device, electronic equipment and readable storage medium |
| CN112614088A (en) * | 2020-12-01 | 2021-04-06 | 安徽维德工业自动化有限公司 | Identification and detection method based on 3D visual detection technology |
| CN113096110A (en) * | 2021-01-15 | 2021-07-09 | 深圳锦绣创视科技有限公司 | Defect self-detection method based on deep learning and related device |
| CN113096110B (en) * | 2021-01-15 | 2024-01-23 | 深圳锦绣创视科技有限公司 | Flaw autonomous detection method based on deep learning and related device |
| CN113252567A (en) * | 2021-06-08 | 2021-08-13 | 菲特（天津）检测技术有限公司 | Method, system, medium and terminal for rapidly detecting multiple defects on surface of aluminum valve plate |
| CN114486732A (en) * | 2021-12-30 | 2022-05-13 | 武汉光谷卓越科技股份有限公司 | Ceramic tile defect online detection method based on line scanning three-dimension |
| CN114486732B (en) * | 2021-12-30 | 2024-04-09 | 武汉光谷卓越科技股份有限公司 | Ceramic tile defect online detection method based on line scanning three-dimension |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108490000A (en) | A kind of Bar Wire Product surface defect on-line measuring device and method | |
| US11830179B2 (en) | Food inspection assisting system, food inspection assisting apparatus and computer program | |
| Momin et al. | Machine vision based soybean quality evaluation | |
| CN102854192B (en) | Method for detecting apple surface defect based on apple surface defect detecting system | |
| Xiong et al. | Color sensors and their applications based on real-time color image segmentation for cyber physical systems | |
| CN106770362A (en) | Gross imperfection detection means and method based on AOI | |
| KR101692115B1 (en) | Inspection apparatus and inspection method | |
| CN103808669B (en) | A kind of apple small holes caused by worms fast non-destructive detection method based on high light spectrum image-forming technology | |
| CN109492714A (en) | Image processing apparatus and its method | |
| Cao et al. | Sparse photometric 3D face reconstruction guided by morphable models | |
| CN207908384U (en) | A kind of double light spectrum image-forming rails and fastener abnormal detector | |
| CN108645814B (en) | Hyperspectral image acquisition method for identifying wetting area of multicolor fabric | |
| CN110441312A (en) | A kind of surface defects of products detection system based on multispectral imaging | |
| JP2000508095A (en) | Boundary mapping system and method | |
| CN106251328B (en) | A method of pollutants on surface of whole chicken rinses is identified using two-dimentional high spectrum image | |
| CN111462128A (en) | Pixel-level image segmentation system and method based on multi-modal spectral image | |
| US20120128232A1 (en) | Method for detecting a bridge connecting failure | |
| CN111812013A (en) | Method for optically detecting biomarkers | |
| CN107833223B (en) | Fruit hyperspectral image segmentation method based on spectral information | |
| CN110097540A (en) | The visible detection method and device of polygon workpeace | |
| CN107576600B (en) | Quick detection method for matcha granularity grade | |
| CN105223208B (en) | A kind of circuit board detecting template and preparation method thereof, circuit board detecting method | |
| CN114445330A (en) | Method and system for detecting appearance defects of components | |
| CN106530315B (en) | Target extraction system and method for medium and small objects under full angle | |
| CN114216867A (en) | Hyperspectral image acquisition and identification device and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180904 |
|
| RJ01 | Rejection of invention patent application after publication |